Source apportionment is critical for guiding development of efficient watershed nitrogen (N) pollution control measures. The ReNuMa (Regional Nutrient Management) model, a semi-empirical, semi-process-oriented model with modest data requirements, has been widely used for riverine N source apportionment. However, the ReNuMa model contains limitations for addressing long-term N dynamics by ignoring temporal changes in atmospheric N deposition rates and N-leaching lag effects. This work modified the ReNuMa model by revising the source code to allow yearly changes in atmospheric N deposition and incorporation of N-leaching lag effects into N transport processes. The appropriate N-leaching lag time was determined from cross-correlation analysis between annual watershed individual N source inputs and riverine N export. Accuracy of the modified ReNuMa model was demonstrated through analysis of a 31-year water quality record (1980–2010) from the Yongan watershed in eastern China. The revisions considerably improved the accuracy (Nash-Sutcliff coefficient increased by ∼0.2) of the modified ReNuMa model for predicting riverine N loads. The modified model explicitly identified annual and seasonal changes in contributions of various N sources (i.e., point vs. nonpoint source, surface runoff vs. groundwater) to riverine N loads as well as the fate of watershed anthropogenic N inputs. Model results were consistent with previously modeled or observed lag time length as well as changes in riverine chloride and nitrate concentrations during the low-flow regime and available N levels in agricultural soils of this watershed. The modified ReNuMa model is applicable for addressing long-term changes in riverine N sources, providing decision-makers with critical information for guiding watershed N pollution control strategies.

Accurate source identification is critical for optimizing water pollution control strategies. Although the dual stable isotope (15N-NO3-/18O-NO3-) approach has been widely applied for differentiating riverine nitrogen (N) sources, the relatively short-term (<1 yr) 15N-NO3-/18O-NO3- records typically used in previous studies often hinders rigorous assessment due to high temporal variability associated with watershed N dynamics. Estimated contributions of legacy N sources in soils and groundwater to riverine N export by modeling approaches in many previous studies also lack validation from complementary information, such as multiple stable isotopes. This study integrated three years of multiple stable isotope (15N-NO3-/18O-NO3- and 2H-H2O/18O-H2O) and hydrochemistry measurements for river water, groundwater and rainfall to elucidate N dynamics and sources in the Yongan watershed (2474 km2) of eastern China. Nonpoint source N pollution dominated and displayed considerable seasonal and spatial variability in N forms and concentrations. Information from δ15N-NO3- and δ18O-NO3- indicated that riverine N dynamics were regulated by contributing sources, nitrification and denitrification, as well as hydrological processes. For the three examined catchments and entire watershed, slow subsurface and groundwater flows accounted for >75% of river discharge and were likely the major hydrological pathways for N delivery to the river. Riverine NO3- sources varied with dominant land use (p < 0.001), with the highest contributions of groundwater (60%), wastewater (35%), and soil (50%) occurring in agricultural, residential and forest catchments, respectively. For the entire watershed, groundwater (∼50%) and soil N (>30%) were the dominant riverine NO3- sources, implying considerable potential for N pollution legacy effects. Results were consistent with observed nitrous oxide dynamics and N sources identified in previous modeling studies. As the first attempt to apply multiple isotope tracers for exploring and quantifying N transformation and transport pathways, this study provides an integrated approach for verifying and understanding the N pollution legacy effects observed in many watersheds worldwide. This study highlights that river N pollution control in many watersheds requires particular attention to groundwater restoration and soil N management in addition to N input control strategies.

Quantitative information on long-term net anthropogenic phosphorus inputs (NAPI) and its relationship with riverine phosphorus (P) export are critical for developing sustainable and efficient watershed P management strategies. This is the first study to address long-term (1980-2015) NAPI and riverine P flux dynamics for the Yangtze River basin (YRB), the largest watershed in China. Over the 36-year study period, estimated NAPI to the YRB progressively increased by ∼1.4 times, with NAPIA (chemical fertilizer input + atmospheric deposition + seed input) and NAPIB (net food/feed imports + non-food input) contributing 65% and 35%, respectively. Higher population, livestock density and agricultural land area were the main drivers of increasing NAPI. Riverine total phosphorus (TP), particulate phosphorus (PP) and suspended sediment (SS) export at Datong hydrological station (downstream station) decreased by 52%, 75% and 75% during 1980-2015, respectively. In contrast, dissolved phosphorus (DP) showed an increase in both concentration (∼7-fold) and its contribution to TP flux (∼16-fold). Different trends in riverine P forms were mainly due to increasing dam/reservoir construction and changes in vegetation/land use and NAPI components. Multiple regression models incorporating NAPIA, NAPIB, dam/reservoir storage capacity and water discharge explained 84% and 92% of the temporal variability in riverine DP and PP fluxes, respectively. Riverine TP flux estimated as the sum of DP and PP fluxes showed high agreement with measured values (R2 = 0.87, NSE = 0.84), indicating strong efficacy for the developed models. The model forecasted an increase of 50% and 7% and a decrease of 15% and 22% in riverine DP flux from 2015 to 2045 under developing, dam building, NAPIA and NAPIB reduction scenarios, respectively. This study highlights the importance of including enhanced P transformation from particulate to bioavailable forms due to river regulation and changes in land-use, input sources and legacy P pools in development of P pollution control strategies.

Incretin-potentiated glucose-stimulated insulin secretion (GSIS) is critical to maintaining euglycemia, of which GLP-1 receptor (GLP-1R) on β-cells plays an indispensable role. Recently, α-cell-derived glucagon but not intestine-derived GLP-1 has been proposed as the critical hormone that potentiates GSIS via GLP-1R. However, the function of glucagon receptors (GCGR) on β-cells remains elusive. Here, using GCGR or GLP-1R antagonists, in combination with glucagon, to treat single β-cells, α-β cell clusters and isolated islets, we found that glucagon potentiates insulin secretion via β-cell GCGR at physiological but not high concentrations of glucose. Furthermore, we transfected primary mouse β-cells with RAB-ICUE (a genetically encoded cAMP fluorescence indicator) to monitor cAMP level after glucose stimulation and GCGR activation. Using specific inhibitors of different adenylyl cyclase (AC) family members, we revealed that high glucose concentration or GCGR activation independently evoked cAMP elevation via AC5 in β-cells, thus high glucose stimulation bypassed GCGR in promoting insulin secretion. Additionally, we generated β-cell-specific GCGR knockout mice which glucose intolerance was more severe when fed a high-fat diet (HFD). We further found that β-cell GCGR activation promoted GSIS more than GLP-1R in HFD, indicating the critical role of GCGR in maintaining glucose homeostasis during nutrient overload.